Amazon EC2 Auto Scaling group capacity providers - Amazon Elastic Container Service
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Amazon EC2 Auto Scaling group capacity providers

When you use Amazon EC2 instances for your capacity, you use Auto Scaling groups to manage the Amazon EC2 instances registered to their clusters. Auto Scaling helps you ensure that you have the correct number of Amazon EC2 instances available to handle the load for your application.

You can use the managed scaling feature to have Amazon ECS manage the scale-in and scale-out actions of the Auto Scaling group (managed scaling) or you can manage the scaling actions yourself. For more information, see Amazon ECS cluster auto scaling.

Auto Scaling group capacity providers considerations

Consider the following when using Auto Scaling group capacity providers in the console:

  • We recommend that you create a new empty Auto Scaling group to use with a capacity provider rather than using an existing one. If you use an existing Auto Scaling group, any Amazon EC2 instances that are associated with the group that were already running and registered to an Amazon ECS cluster before the Auto Scaling group being used to create a capacity provider might not be properly registered with the capacity provider. This might cause issues when using the capacity provider in a capacity provider strategy. Use DescribeContainerInstances to confirm whether a container instance is associated with a capacity provider or not.

    Note

    To create an empty Auto Scaling group, set the desired count to zero. After you created the capacity provider and associated it with a cluster, you can then scale it out.

    When you use the Amazon ECS console Create Cluster with the Amazon EC2 instances option under Infrastructure, Amazon ECS creates an Amazon EC2 Auto Scaling launch configuration and Auto Scaling group on your behalf as part of the Amazon CloudFormation stack. They are prefixed with EC2ContainerService-<ClusterName>, which make them easy to identify. You can use the Auto Scaling group as a capacity provider for that cluster.

  • An Auto Scaling group must have a MaxSize greater than zero to scale out.

  • The Auto Scaling group can't have instance weighting settings. Instance weighting isn't supported when used with an Amazon ECS capacity provider.

  • If the Auto Scaling group can't scale out to accommodate the number of tasks run, the tasks fails to transition beyond the PROVISIONING state.

  • Don't modify the scaling policy resource associated with your Auto Scaling groups that are managed by capacity providers.

  • When you use managed termination protection, you must also use managed scaling. Otherwise, managed termination protection won't work.

  • When managed scaling is turned on, the Auto Scaling group capacity provider creates a scaling policy resource to manage the scaling of your Auto Scaling group. You can identify these resources by the ECSManaged prefix.

    When you use managed termination protection, Amazon ECS only terminates EC2 instances that don't have any running Amazon ECS tasks.

  • If managed termination protection is turned on when you create a capacity provider, the Auto Scaling group and each Amazon EC2 instance in the Auto Scaling group must have instance protection from scale in turned on. For more information, see Instance Protection in the Amazon Auto Scaling User Guide.

  • You can add a warm pool to your Auto Scaling group. A warm pool is a group of pre-initialized Amazon EC2 instances that are ready to be included in the cluster whenever your application needs to scale out. For more information about warm pools, see Using a warm pool for your Auto Scaling group.

  • If managed scaling is turned on when you create a capacity provider, the Auto Scaling group desired count can be set to 0. When managed scaling is turned on, Amazon ECS manages the scale-in and scale-out actions of the Auto Scaling group.

  • Managed instance draining is turned on by default when you create a capacity provider. We recommend you use this feature to enable graceful termination of Amazon EC2 instances without disrupting your workloads.

For more information about creating an Amazon EC2 Auto Scaling launch template, see Launch Templates in the Amazon EC2 Auto Scaling User Guide. For more information about creating an Amazon EC2 Auto Scaling group, see Auto Scaling groups in the Amazon EC2 Auto Scaling User Guide.

Using a warm pool for your Auto Scaling group

Amazon ECS supports Amazon EC2 Auto Scaling warm pools. A warm pool is a group of pre-initialized Amazon EC2 instances ready to be placed into service. Whenever your application needs to scale out, Amazon EC2 Auto Scaling uses the pre-initialized instances from the warm pool rather than launching cold instances, allows for any final initialization process to run, and then places the instance into service.

To learn more about warm pools and how to add a warm pool to your Auto Scaling group, see Warm pools for Amazon EC2 Auto Scaling in the Amazon EC2 Auto Scaling User Guide.

When you create or update a warm pool for an Auto Scaling group for Amazon ECS , you cannot set the option that returns instances to the warm pool on scale in (ReuseOnScaleIn). For more information, see put-warm-pool in the Amazon Command Line Interface Reference.

To use warm pools with your Amazon ECS cluster, set the ECS_WARM_POOLS_CHECK agent configuration variable to true in the User data field of your Amazon EC2 Auto Scaling group launch template. The following shows an example of how the agent configuration variable can be specified in the User data field of an Amazon EC2 launch template.

#!/bin/bash cat <<'EOF' >> /etc/ecs/ecs.config ECS_CLUSTER=MyCluster ECS_WARM_POOLS_CHECK=true EOF

The ECS_WARM_POOLS_CHECK variable is only supported on agent versions 1.59.0 and later. For more information about the variable, see Amazon ECS container agent configuration.